What software did you use for this?

Can you explain why you took sampling time as 0.5ms?

Magnolia Cove

Real Ultrasonics sensors has few additional parameters 1) typical sampling frequency is 50Hz so they won't update in real time. If robot moves with speed 1m/s toward wall, it moves 2cm before update. It's even more significant during rotation. 2) ultasonic sensors has high noise. Usually +/- 2cm for cheap one with 250cm range. 3) They can have issue... Same model of US usually works on same frequency = you has to trigger sensors one by one because they interfere with each other.

7:20 F-Bomb dropped

Will we see future videos of ROS and Gazebo using deep reinforcement learning?

hi i really want to know how to i can create model 'h5'

Is there a place to play these

Enjoyed this content ?

Hello. I implemented a state feedback control using the state-space model with the provided A, B, C, and D matrices in Simulink, but I noticed that the state variables are diverging. When I checked the poles using the eig function, I found that there is a positive real number included. Why does the regulator work correctly in the Simscape model, but when I create a separate state feedback model, it produces diverging values? Also, when designing a state feedback model in Simulink, should the initial value of the integrator be set to the same value as the step input in Simscape?

ROBOT.py----> import pygame import math import numpy as np def distance(point1,point2): point1 = np.array(point1) point2 = np.array(point2) return np.linalg.norm(point1- point2) class Robot: def __init__(self, startpos,width): self.m2p= 3779.52 #from meters to pixels #robot dims self.w = width self.x= startpos[0] self.y = startpos[1] self.heading =0 self.vl = 0.01*self.m2p #meters/s self.vr = 0.01*self.m2p self.maxspeed = 0.02*self.m2p self.minspeed = 0.01*self.m2p self.min_obs_dist = 100 self.count_down = 5 #seconds def avoid_obstacle(self,point_cloud,dt): closest_obs = None dist = np.inf if len(point_cloud) > 1: for point in point_cloud: if dist > distance([self.x, self.y], point): dist = distance([self.x,self.y], point) closest_obs = (point,dist) if closest_obs[1] < self.min_obs_dist and self.count_down > 0: self.count_down -= dt self.move_backward() else: #reset count down self.count_down = 5 #move forward self.move_forward() def move_backward(self): self.vr = - self.minspeed self.vl = - self.minspeed/2 def move_forward(self): self.vr = self.minspeed self.vl = self.minspeed def kinematics(self, dt): self.x += ((self.vl+self.vr)/2) * math.cos(self.heading) * dt self.y -= ((self.vl+self.vr)/2) * math.sin(self.heading) * dt self.heading += (self.vr - self.vl) / self.w * dt if self.heading>2*math.pi or self.heading<-2*math.pi: self.heading = 0 self.vr = max(min(self.maxspeed, self.vr), self.minspeed) self.vl = max(min(self.maxspeed, self.vr), self.minspeed) class Graphics: def __init__(self, dimentions, robot_img_path, map_img_path): pygame.init() #COLORS self.black = (0, 0, 0) self.white =(255, 255, 255) self.green =(0, 255, 0) self.blue =(0, 0, 255) self.red = (255, 0, 0) self.yel = (255, 255, 0) #----MAP---- #load imgs self.robot = pygame.image.load(robot_img_path) self.map_img = pygame.image.load(map_img_path) #dimentions self.height, self.width = dimentions #window settings pygame.display.set_caption("Obstacle Avoidance") self.map = pygame.display.set_mode((self.width, self.height)) self.map.blit(self.map_img, (0,0)) def draw_robot(self,x,y, heading): rotated = pygame.transform.rotozoom(self.robot, math.degrees(heading), 1) rect = rotated.get_rect(center=(x, y)) self.map.blit(rotated, rect) def draw_sensor_data(self, point_cloud): for point in point_cloud: pygame.draw.circle(self.map, self.red, point, 3, 0) class Ultrasonic: def __init__(self, sensor_range, map): self.sensor_range = sensor_range self.map_width, self.map_height = pygame.display.get_surface().get_size() self.map = map def sense_obstacles(self, x, y, heading): obstacles = [] x1, y1 = x, y start_angle = heading - self.sensor_range[1] finish_angle = heading + self.sensor_range[1] for angle in np.linspace(start_angle, finish_angle , 10, False): x2 = x1 + self.sensor_range[0] * math.cos(angle) y2 = y1 - self.sensor_range[0] * math.sin(angle) for i in range(0, 100): u = i / 100 x= int(x2 * u + x1 * (1-u )) y= int(y2 * u + y1 * (1-u )) if 0 < x < self.map_width and 0 < y < self.map_height: color = self.map.get_at((x,y)) self.map.set_at((x,y), (0, 208,255)) if (color[0], color[1], color[2]) == (0,0,0): obstacles.append([x,y]) break return obstacles main.py --> import math import pygame from ROBOT import Graphics, Robot, Ultrasonic MAP_DIMENSIONS = (600, 1200) #the environment graphics gfx = Graphics(MAP_DIMENSIONS, 'D:\college\internship\INTERNSHIP TASK\obstacle detection\images\DDR.png', 'D:\college\internship\INTERNSHIP TASK\obstacle detection\images\Obstacle.png') #the robot start = (200, 200) robot = Robot(start, 0.01*3779.52) #the sensor sensor_range = 250, math.radians(40) ultra_sonic = Ultrasonic(sensor_range, gfx.map) dt=0 last_time = pygame.time.get_ticks() running = True #simulation loop while running: for event in pygame.event.get(): if event.type == pygame.QUIT: running = False dt= (pygame.time.get_ticks()-last_time)/1000 last_time = pygame.time.get_ticks() gfx.map.blit(gfx.map_img, (0,0)) robot.kinematics(dt) gfx.draw_robot(robot.x, robot.y, robot.heading) point_cloud = ultra_sonic.sense_obstacles(robot.x, robot.y, robot.heading) robot.avoid_obstacles(point_cloud, dt) gfx.draw_sensor_data(point_cloud) pygame.display.update THANK ME LATER😌

Thank you Sir for the video. I tried to run the code in the video install scipy version does not work

Hello, are your services still available on fiverr?

🔗Source Code: Rubik's Cube simulation: ko-fi.com/s/8558c6b0a0 Rubik's Cube simulation (machine learning solver) : Coming soon

That's cool bro but you mispositioned the colors of the Rubik cube

That's a really nice project and your explanation was easy to follow keep up the good work 👍

That's cool❤

How did you like this second part Rubik's Cube simulation: ko-fi.com/s/8558c6b0a0 Rubik's Cube simulation (machine learning solver) : Coming soon

I really liked this tutorial, I’m looking forward for part 2

que genio

I just took a look into your older videos - you do have a good narrating voice, please use it instead of this annoying TTS/AI voice

Impressive as usual Thank you for sharing your work with us

Hello i'm back 🔗Source Code: Rubik's Cube simulation: ko-fi.com/s/8558c6b0a0 Rubik's Cube simulation (machine learning solver) : Coming soon

the code executed fine and the output video file was generated. but when i played the video the lanes weren't marked with green color, or highlighted in anyway. It looked same as the input video. where could i be missed something ? i followed your code line by line . please help

can this be use in Lightbulbs?

interesting video although just doing a standard screen cast even if it's longer would be much more useful. I feel like i have to debug this tutorial. You add in blocks and components offscreen and use keyboard short cuts without explaining how to search for components. Also things like at 12:12 where you say it's really crucial to set a value and then cover the actual screen with a blown up screen shot of a properties window (but is it the one you're talking about, is it a bug and which one is it for), below the screenshot you can see the bottom of the actual window you're editing.. Great topic but honestly just a screen share of the whole process (sped up for bits if needed) would be much more useful. Nothing fancy just what you did is best.

Hello great video sir, i was wondering how do you suppose is it possible using the same model for the robot to make turns to reach a certain target rather than just move forward and backward? An idea or resource would be highly appreciated. Thank you.

can i ask is this work in light bulb?

i have a drim to be a youtuber like you keep going bro

in this simulation you are just calculation vertical angle?

Is it more performant to find the intersection between (x1,y1),(x2,y2) and the four sides of the rectangle instead of splitting it to 100 points?

Great videos loving it please keep making these videos

hey is there any way of doing this with camera but without a LIDAR? because I've done some research and it kinda looks like it's possible...

📁 source code: ko-fi.com/s/b0f376134e

Good video .How to convert signal position of revolute to voltage to be inserted to the pid to get a force acting on prismatic

So one good collie 👍

Python Sheep Herding Simulation 🗃️ source code: ko-fi.com/s/f5c64c0476

Hello, can you upload a video of adding a PID controller to the robot you imported from onshape. Like a walking robot.. thank you

Thank you very much dear sir. You teach excellent but please describe more about something like 100 signals or sample time! I exactly make your model . My motor turns about 3200 rpm but the result is not acceptable.

🗃️ source code : ko-fi.com/s/1a38e1563b

I use rplidar c1.. when i run it apper a black window and the lidar doesn't connect or even work!😢 please help me

🏷️ Download the files: ko-fi.com/s/716bba8384

code: ko-fi.com/s/86d7053723

If you did all right and still having and error try this: "File" -> "Invalidate Caches / Restart" -> "Invalidate and Restart". This restart the old cache and validates the added modules like: import pygame etc...

Hello! I wrote all the code like yours, but it doesn't display, scan, or display an image.

most part of the main.py code didnt show in the later part of the video

Hey are you still on fiverr available?

I still didn't understand the purpose of adding delay blocks. Z-1 and Z-100